i18npool/source/characterclassification/scripttypedetector.cxx

   1 /*************************************************************************
   2  *
   3  * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
   4  *
   5  * Copyright 2008 by Sun Microsystems, Inc.
   6  *
   7  * OpenOffice.org - a multi-platform office productivity suite
   8  *
   9  * $RCSfile: scripttypedetector.cxx,v $
  10  * $Revision: 1.10 $
  11  *
  12  * This file is part of OpenOffice.org.
  13  *
  14  * OpenOffice.org is free software: you can redistribute it and/or modify
  15  * it under the terms of the GNU Lesser General Public License version 3
  16  * only, as published by the Free Software Foundation.
  17  *
  18  * OpenOffice.org is distributed in the hope that it will be useful,
  19  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  20  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  21  * GNU Lesser General Public License version 3 for more details
  22  * (a copy is included in the LICENSE file that accompanied this code).
  23  *
  24  * You should have received a copy of the GNU Lesser General Public License
  25  * version 3 along with OpenOffice.org.  If not, see
  26  * <http://www.openoffice.org/license.html>
  27  * for a copy of the LGPLv3 License.
  28  *
  29  ************************************************************************/
  30
  31 // MARKER(update_precomp.py): autogen include statement, do not remove
  32 #include "precompiled_i18npool.hxx"
  33
  34 #include <com/sun/star/i18n/CTLScriptType.hpp>
  35 #include <com/sun/star/i18n/ScriptDirection.hpp>
  36 #include <com/sun/star/i18n/UnicodeScript.hpp>
  37 #include <scripttypedetector.hxx>
  38 #include <i18nutil/unicode.hxx>
  39
  40 //      ----------------------------------------------------
  41 //      class ScriptTypeDetector
  42 //      ----------------------------------------------------;
  43
  44 using namespace com::sun::star::i18n;
  45
  46 ScriptTypeDetector::ScriptTypeDetector()
  47 {
  48 }
  49
  50 ScriptTypeDetector::~ScriptTypeDetector()
  51 {
  52 }
  53
  54 static sal_Int16 scriptDirection[] = {
  55     ScriptDirection::LEFT_TO_RIGHT,     // DirectionProperty_LEFT_TO_RIGHT = 0,
  56     ScriptDirection::RIGHT_TO_LEFT,     // DirectionProperty_RIGHT_TO_LEFT = 1,
  57     ScriptDirection::LEFT_TO_RIGHT,     // DirectionProperty_EUROPEAN_NUMBER = 2,
  58     ScriptDirection::LEFT_TO_RIGHT,     // DirectionProperty_EUROPEAN_NUMBER_SEPARATOR = 3,
  59     ScriptDirection::LEFT_TO_RIGHT,     // DirectionProperty_EUROPEAN_NUMBER_TERMINATOR = 4,
  60     ScriptDirection::RIGHT_TO_LEFT,     // DirectionProperty_ARABIC_NUMBER = 5,
  61     ScriptDirection::NEUTRAL,           // DirectionProperty_COMMON_NUMBER_SEPARATOR = 6,
  62     ScriptDirection::NEUTRAL,           // DirectionProperty_BLOCK_SEPARATOR = 7,
  63     ScriptDirection::NEUTRAL,           // DirectionProperty_SEGMENT_SEPARATOR = 8,
  64     ScriptDirection::NEUTRAL,           // DirectionProperty_WHITE_SPACE_NEUTRAL = 9,
  65     ScriptDirection::NEUTRAL,           // DirectionProperty_OTHER_NEUTRAL = 10,
  66     ScriptDirection::LEFT_TO_RIGHT,     // DirectionProperty_LEFT_TO_RIGHT_EMBEDDING = 11,
  67     ScriptDirection::LEFT_TO_RIGHT,     // DirectionProperty_LEFT_TO_RIGHT_OVERRIDE = 12,
  68     ScriptDirection::RIGHT_TO_LEFT,     // DirectionProperty_RIGHT_TO_LEFT_ARABIC = 13,
  69     ScriptDirection::RIGHT_TO_LEFT,     // DirectionProperty_RIGHT_TO_LEFT_EMBEDDING = 14,
  70     ScriptDirection::RIGHT_TO_LEFT,     // DirectionProperty_RIGHT_TO_LEFT_OVERRIDE = 15,
  71     ScriptDirection::NEUTRAL,           // DirectionProperty_POP_DIRECTIONAL_FORMAT = 16,
  72     ScriptDirection::NEUTRAL,           // DirectionProperty_DIR_NON_SPACING_MARK = 17,
  73     ScriptDirection::NEUTRAL,           // DirectionProperty_BOUNDARY_NEUTRAL = 18,
  74 };
  75
  76 sal_Int16 SAL_CALL
  77 ScriptTypeDetector::getScriptDirection( const ::rtl::OUString& Text, sal_Int32 nPos, sal_Int16 defaultScriptDirection ) throw (::com::sun::star::uno::RuntimeException)
  78 {
  79     sal_Int16 dir = scriptDirection[unicode::getUnicodeDirection(Text[nPos])];
  80     return (dir == ScriptDirection::NEUTRAL) ? defaultScriptDirection : dir;
  81 }
  82
  83 // return value '-1' means either the direction on nPos is not same as scriptDirection or nPos is out of range.
  84 sal_Int32 SAL_CALL
  85 ScriptTypeDetector::beginOfScriptDirection( const ::rtl::OUString& Text, sal_Int32 nPos, sal_Int16 direction ) throw (::com::sun::star::uno::RuntimeException)
  86 {
  87         sal_Int32 cPos = nPos;
  88
  89         if (cPos < Text.getLength()) {
  90             for (; cPos >= 0; cPos--) {
  91                 if (direction != getScriptDirection(Text, cPos, direction))
  92                     break;
  93             }
  94         }
  95         return cPos == nPos ? -1 : cPos + 1;
  96 }
  97
  98 sal_Int32 SAL_CALL
  99 ScriptTypeDetector::endOfScriptDirection( const ::rtl::OUString& Text, sal_Int32 nPos, sal_Int16 direction ) throw (::com::sun::star::uno::RuntimeException)
 100 {
 101         sal_Int32 cPos = nPos;
 102         sal_Int32 len = Text.getLength();
 103
 104         if (cPos >=0) {
 105             for (; cPos < len; cPos++) {
 106                 if (direction != getScriptDirection(Text, cPos, direction))
 107                     break;
 108             }
 109         }
 110         return cPos == nPos ? -1 : cPos;
 111 }
 112
 113 sal_Int16 SAL_CALL
 114 ScriptTypeDetector::getCTLScriptType( const ::rtl::OUString& Text, sal_Int32 nPos ) throw (::com::sun::star::uno::RuntimeException)
 115 {
 116     static ScriptTypeList typeList[] = {
 117         { UnicodeScript_kHebrew, UnicodeScript_kHebrew, CTLScriptType::CTL_HEBREW },    // 10
 118         { UnicodeScript_kArabic, UnicodeScript_kArabic, CTLScriptType::CTL_ARABIC },    // 11
 119         { UnicodeScript_kDevanagari, UnicodeScript_kDevanagari, CTLScriptType::CTL_INDIC },     // 14
 120         { UnicodeScript_kThai, UnicodeScript_kThai, CTLScriptType::CTL_THAI },      // 24
 121         { UnicodeScript_kScriptCount, UnicodeScript_kScriptCount,   CTLScriptType::CTL_UNKNOWN }    // 88
 122     };
 123
 124     return unicode::getUnicodeScriptType(Text[nPos], typeList, CTLScriptType::CTL_UNKNOWN);
 125 }
 126
 127 // Begin of Script Type is inclusive.
 128 sal_Int32 SAL_CALL
 129 ScriptTypeDetector::beginOfCTLScriptType( const ::rtl::OUString& Text, sal_Int32 nPos ) throw (::com::sun::star::uno::RuntimeException)
 130 {
 131     if (nPos < 0)
 132         return 0;
 133     else if (nPos >= Text.getLength())
 134         return Text.getLength();
 135     else {
 136         sal_Int16 cType = getCTLScriptType(Text, nPos);
 137         for (nPos--; nPos >= 0; nPos--) {
 138             if (cType != getCTLScriptType(Text, nPos))
 139                 break;
 140         }
 141         return nPos + 1;
 142     }
 143 }
 144
 145 // End of the Script Type is exclusive, the return value pointing to the begin of next script type
 146 sal_Int32 SAL_CALL
 147 ScriptTypeDetector::endOfCTLScriptType( const ::rtl::OUString& Text, sal_Int32 nPos ) throw (::com::sun::star::uno::RuntimeException)
 148 {
 149     if (nPos < 0)
 150         return 0;
 151     else if (nPos >= Text.getLength())
 152         return Text.getLength();
 153     else {
 154         sal_Int16 cType = getCTLScriptType(Text, nPos);
 155         sal_Int32 len = Text.getLength();
 156         for (nPos++; nPos < len; nPos++) {
 157             if (cType != getCTLScriptType(Text, nPos))
 158                 break;
 159         }
 160         return nPos;
 161     }
 162 }
 163
 164 const sal_Char sDetector[] = "draft.com.sun.star.i18n.ScriptTypeDetector";
 165
 166 rtl::OUString SAL_CALL
 167 ScriptTypeDetector::getImplementationName() throw( ::com::sun::star::uno::RuntimeException )
 168 {
 169     return ::rtl::OUString::createFromAscii(sDetector);
 170 }
 171
 172 sal_Bool SAL_CALL
 173 ScriptTypeDetector::supportsService(const rtl::OUString& ServiceName) throw( ::com::sun::star::uno::RuntimeException )
 174 {
 175     return !ServiceName.compareToAscii(sDetector);
 176 }
 177
 178 ::com::sun::star::uno::Sequence< rtl::OUString > SAL_CALL
 179 ScriptTypeDetector::getSupportedServiceNames() throw( ::com::sun::star::uno::RuntimeException )
 180 {
 181     ::com::sun::star::uno::Sequence< ::rtl::OUString > aRet(1);
 182     aRet[0] = ::rtl::OUString::createFromAscii(sDetector);
 183     return aRet;
 184 }
 185